IPM Take
Emergency shelter is a health service, whether planners call it one or not.
Beds, food and physical safety are essential. They are not the full operating model. Crowded shelters bring together displaced residents, rotating staff and multiple agencies while normal health records, routines and privacy are disrupted. Infection prevention has to be built into the opening plan, with named responsibilities and supplies already in place.
Executive Summary
A CDC MMWR investigation documented 104 norovirus cases, 56 COVID-19 cases, 29 influenza cases and 30 nonspecific respiratory illnesses among residents and staff at the Eaton wildfire evacuation shelter in Pasadena in January-February 2025. Nine people with norovirus and six with COVID-19 were hospitalised. Public-health teams introduced designated isolation areas, off-site isolation for COVID-19, enhanced surveillance, cleaning protocols, staff protections and regular infection-control audits. The report concluded that early planning, shared priorities and consistent protocols across agencies were central to controlling the outbreaks.
Why it matters
- Emergency planners: Need infection-prevention roles, isolation space, supplies and reporting protocols embedded in shelter activation plans.
- Health providers: Need on-site clinical pathways that protect privacy, identify severe illness and maintain continuity during agency handovers.
- Shelter residents and staff: Need safe facilities and clear information without being blamed for infections created by crowded emergency conditions.
The wildfire forced people from their homes. The shelter then had to manage three outbreaks at once.
That sequence is not an argument against evacuation shelters. It is an argument against treating them as logistics sites with health care added later. The Eaton shelter was a temporary living environment, workplace, clinic and outbreak setting at the same time.
CDC’s investigation identified norovirus, COVID-19 and influenza among residents and staff, alongside other respiratory illness. Some patients required hospital care. The case totals were probably underestimates because surveillance depended in part on people seeking care and staff monitoring began later.
The response improved when responsibilities became explicit. Public-health agencies and partner organisations designated isolation areas, strengthened illness reporting, reinforced hand hygiene and environmental cleaning, supplied appropriate disinfectants, monitored mask use and conducted regular audits. COVID-19 cases were isolated off site; people with gastrointestinal or other respiratory illness used separate areas and facilities.
The operational lesson is less glamorous than a new technology and more important than most emergency plans admit: protocols must survive staff turnover and agency transitions. A shelter may be run by different organisations over several weeks. If illness definitions, reporting routes or cleaning standards change with every handover, the outbreak gains an advantage each time management changes.
Privacy matters as well. Surveillance that includes staff and residents can improve containment, but only when personal information is protected and people do not fear losing shelter, work or dignity by reporting symptoms.
Climate-related displacement will make this problem harder. Heat, fires, floods and storms can create crowded temporary settings precisely when local clinics, laboratories and public-health offices are also disrupted. Preparedness must therefore include portable infection-control capacity, not assume the normal health system will remain fully available.
For IPM, the Eaton experience turns a retrospective report into a forward requirement. When the next shelter opens, infection prevention should already be in the floor plan, staffing model and command structure. Waiting for the first cluster is not flexibility. It is preventable delay.
